JP2016071591A - Verification device and verification method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the overall processing time required for collation, while improving collation accuracy.SOLUTION: A collation apparatus (1) includes: an image data acquisition section (10) for acquiring image data (21); a motion detection processing section (12) which detects whether a subject included in the image data has moved or not, on the basis of the acquired image data; a collation image data generation section (13) which generates collation data to be used for collation, on the basis of the acquired image data, when motion of the subject is not detected, and does not generate collation image data when a motion of the subject is detected; a storage section (15) which stores registration image data (151_1-151_n) formed by imaging the subject to be collated; and a collation processing section (16) which collates the collation image data generated by the collation data generation section with the registration image data stored in the storage section.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a collation apparatus and a collation method, for example, a collation apparatus that performs collation based on a captured image for face authentication, fingerprint authentication, and the like.

  In recent years, in areas where personal authentication is required, such as entry / exit management for computer rooms and critical machine rooms, and access control for computer terminals and financial terminals, fingerprint authentication and face cards have been used instead of personal identification numbers and ID cards. Biometric authentication such as authentication is being adopted. For example, Patent Document 1 discloses a collation device that performs identity verification by collating pre-registered face or fingerprint image data with face or fingerprint image data captured by a camera or sensor. Yes.

  In such a collation device, when a photographed image is blurred, there is a possibility that it will adversely affect the processes relating to the subsequent collation and image registration. For example, when collating image data such as a captured face or fingerprint with registered image data, if the image used for collation is blurred, the person collation may fail or it may be misidentified as another person. Increases and collation accuracy decreases. Further, if a blurred image is registered as it is when registering image data such as a face and a fingerprint, an error may occur during subsequent verification, and the verification accuracy may be lowered. Therefore, it is necessary to re-take a registered image.

  As a conventional technique for avoiding blurring of an image at the time of shooting, for example, Patent Document 2 discloses an imaging device that determines shooting timing based on a change in the state of a subject. In the imaging apparatus disclosed in Patent Literature 2, it is determined whether or not the subject is stationary based on the face detection results of two or more images acquired by performing pre-photographing before the main photographing at the time of timer photographing. However, the main shooting is performed after determining that the camera is stationary.

Japanese Patent Application Laid-Open No. 09-022406 JP 2006-254358 A

However, the technique described in Patent Document 2 determines that the subject is stationary based on an image acquired by pre-shooting, and determines whether the image acquired by main shooting is blurred. Rather, there is no guarantee that the image from the actual shooting is not blurred. For this reason, even if the technique described in Patent Document 2 is applied to a collation device, it cannot be said that the above-described adverse effects on the processes related to collation and image registration can be prevented appropriately.
In addition, in the conventional collation device, since the collation process is performed regardless of whether or not the captured image is blurred, there is a high possibility that a wasteful collation process is performed when the image is blurred. There has been a problem that the overall processing time required for verification becomes long.

  The present invention has been made to solve the above problems, and an object of the present invention is to improve collation accuracy and shorten the overall processing time required for collation in a collation apparatus.

  The collation device (1) according to the present invention includes an image data acquisition unit (10) that acquires image data (21) of an image captured by the imaging device (2), and image data acquired by the image data acquisition unit. Based on the motion detection processing unit (12) that detects the presence or absence of motion of the subject included in the image data, and when the motion detection processing unit does not detect the motion of the subject, the image data acquisition unit Based on the acquired image data, collation data used for collation is generated, and the collation image data generation unit (13) that does not generate the collation image data when the motion detection processing unit detects the movement of the subject. ), Registered image data (151_1 to 151_n) showing the subject to be verified, and the verification data generation unit Wherein the collation image data generated Te, the verification processing unit for matching with the registered image data stored in the storage unit (16), characterized by having a.

  The collation apparatus may further include an image registration unit (14) that stores the collation image data generated by the collation data generation unit in the storage unit as the registered image data.

  In the verification device, the motion detection processing unit calculates a movement amount and correlation strength of the two image data based on two image data having a photographing time difference acquired by the image data acquisition unit by a phase only correlation method. Then, the presence or absence of the movement of the subject may be detected based on the calculated movement amount and the correlation strength.

  The collation apparatus further includes an area detection unit (11) that detects an area including a characteristic portion of the subject included in the image data acquired by the image data acquisition unit, and the motion detection processing unit includes the 2 Data in the area detected by the area detection unit may be extracted from one image data, and the amount of movement and the correlation strength between the two image data may be calculated based on the extracted data.

  In the verification apparatus, the two image data may be odd frame data and even frame data of an interlaced image.

  The collation apparatus may further include a notification unit (17) for notifying an error and instructing the imaging device to perform re-shooting when the movement of the subject is detected by the motion detection processing unit. Good.

  The collation method according to the present invention includes a first step (S1) for acquiring image data of an image captured by an imaging device, and a subject included in the image data based on the image data acquired by the image data acquisition unit. A second step (S2) for detecting the presence or absence of movement of the subject, and a collation used for the collation based on the image data acquired by the image data acquisition unit when the movement of the subject is not detected in the second step The third step (S5, S6) for generating data, the collation image data generated in the third step, and the registered image data showing the subject to be collated stored in the storage unit are collated. And a fourth step (S7).

  In the above description, the reference numerals with parentheses merely exemplify what are included in the concept of the constituent elements with the reference numerals in the drawings.

  As described above, according to the present invention, the collation apparatus can improve collation accuracy and shorten the overall processing time required for collation.

FIG. 1 is a diagram illustrating a configuration of an authentication system provided with a collation device according to an embodiment of the present invention. FIG. 2 is a diagram showing an overall processing flow of the collating operation by the collating device according to the present embodiment. FIG. 3 is a diagram showing a processing flow of motion detection processing by the collation device according to the present embodiment. FIG. 4 is a diagram illustrating an example of extraction of feature area image data by the motion detection processing unit. FIG. 5A is a diagram showing a correlation characteristic between two pieces of image data by the phase-only correlation method. FIG. 5B is a diagram showing another correlation characteristic between two pieces of image data by the phase-only correlation method. FIG. 6 is a diagram showing a processing flow of collation processing by the collation device according to the present embodiment. FIG. 7 is a diagram showing an overall processing flow of the image registration operation by the collation apparatus according to the present embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[Configuration of verification device]
FIG. 1 is a diagram illustrating a configuration of an authentication system provided with a collation device according to an embodiment of the present invention. An authentication system 100 shown in the figure performs personal verification using biometric authentication such as fingerprint authentication, face authentication, vein authentication, and the like.

  The authentication system 100 includes a verification device 1 and an imaging device 2. The imaging device 2 generates image data 21 for biometric authentication of a person who performs personal authentication. The imaging device 2 is a stereo camera that captures, for example, near-infrared light and visible light, and the generated image data 21 includes 2D image data and 3D image data captured by the stereo camera. The verification device 1 performs various authentication processes based on the image data acquired from the imaging device 2.

  In the present embodiment, as an example, it is assumed that the authentication system 100 performs face authentication, and the matching device 1 performs face authentication based on the image data 21 of an image obtained by photographing the face of the subject by the imaging device 2. A description will be given assuming that various processes are performed.

  Specifically, the collation device 1 is registered with an image registration function for registering a face image of a subject photographed by the imaging device 2 as a collation target image, and a face image photographed by the imaging device 2. And a function of collating with an image to be collated. Furthermore, the collation device 1 has a motion detection function that detects the presence or absence of motion of the subject (presence or absence of blurring of the image) in the image captured by the imaging device 2, and the motion of the subject is not detected. In the case of registration, image registration and verification are performed. Hereinafter, a specific configuration of the verification device 1 will be described.

  The collation apparatus 1 includes an image data acquisition unit 10, an area detection unit 11, a motion detection processing unit 12, a collation data generation unit 13, an image registration unit 14, a storage unit 15, a collation processing unit 16, and a notification unit 17. Yes.

  The image data acquisition unit 10 is a functional unit that acquires image data for biometric authentication of a subject (person) who performs personal authentication from the outside. For example, the image data acquisition unit 10 receives the image data 21 generated by the imaging device 2 and temporarily stores it.

  The area detection unit 11 detects an area including a feature portion included in the image data 21 acquired by the image data acquisition unit 10. Specifically, the area detection unit 11 detects an area in which the “face” of the subject in the image data 21 is captured (hereinafter referred to as “face area”). In addition, the area detection unit 11 detects a feature area specified by a facial feature (for example, a pupil, a nose, and a mouth) from a face image in the image data 21. Moreover, the area detection part 11 produces | generates the area information which shows the detected area. The area information includes, for example, coordinate information indicating coordinates in the image data of the detected area.

  Based on the image data 21 acquired by the image data acquisition unit 10, the motion detection processing unit 12 performs a motion detection process for detecting the presence or absence of movement of the subject included in the image data 21. Specifically, the motion detection processing unit 12 uses the phase-only correlation method to move the two image data 21_1 and 21_2 based on the two image data 21_1 and 21_2 acquired by the image data acquisition unit 10 and having a photographing time difference. dx, dy and correlation strength a are calculated, and based on the calculated movement amounts dx, dy and correlation strength a, the presence or absence of movement of the subject is detected. For example, the motion detection processing unit 12 extracts the image data of the areas specified by the area information (coordinate information) generated by the area detection unit 11 from the two image data acquired by the image data acquisition unit 10, respectively. Based on the two extracted image data, the movement amount and correlation strength between the image data 21_1 and 21_2 are calculated.

  The verification data generation unit 13 generates verification data based on the image data 21 acquired by the image data acquisition unit 10. Here, the collation data is data used in collation processing by the collation processing unit 16 to be described later, and includes, for example, point cloud data representing the shape of the face portion of a person in three dimensions.

  Specifically, the collation data generation unit 13 performs three-dimensional measurement based on the two-dimensional image data and the three-dimensional image data included in the image data 21 when the movement of the subject is not detected by the movement detection processing unit 12. In addition, verification data is generated based on the three-dimensional measurement result. On the other hand, when the movement of the subject is detected by the motion detection processing unit 12, the verification data generation unit 13 does not generate verification image data.

  When the motion detection processing unit 12 detects the movement of the subject, the notification unit 17 notifies an error and instructs the imaging device 2 to perform re-shooting. As a specific error notification method, for example, an error is displayed on a screen of a console (for example, a screen of a touch panel or a liquid crystal monitor) installed in the vicinity of the imaging device 2, so that the user (face authentication target person) is notified. An error and re-shooting may be notified, or an error or the like may be notified to the user by outputting sound from a speaker of the console.

  The storage unit 15 stores biometric authentication image data (hereinafter referred to as “registered image data”) 151_1 to 151_n of n persons (n is an integer of 1 or more) to be collated. The registered image data 151_1 to 151_n are stored in the storage unit 15 in association with the reference number, for example. Note that the registered image data 151_1 to 151_n are collectively referred to as registered image data 151.

  The image registration unit 14 stores the collation data generated by the collation data generation unit 13 in the storage unit 15 as registered image data 151 at the time of image registration.

  The collation processing unit 16 performs collation processing for collating the collation data generated by the collation data generating unit 13 with the registered image data 151 stored in the storage unit 15 at the time of collation. A determination result as to whether or not the image data 151 matches is output.

  The collation device 1 described above includes an arithmetic device such as a CPU, a memory such as a RAM (Random Access Memory) and a ROM, a storage device such as an HDD (Hard Disc Drive), a keyboard, a mouse, a pointing device, an operation button, a touch panel, and the like. An input device that detects the input of information from outside, an I / F device that transmits and receives various information via a communication line such as the Internet, a LAN (Local Area Network), a WAN (Wide Area Network), and an LCD (Liquid) This is realized by a computer having a display device such as Crystal Display) and a program installed in the computer. That is, when the above hardware resources are controlled by a program, the above-described image data acquisition unit 10, area detection unit 11, motion detection processing unit 12, verification data generation unit 13, image registration unit 14, storage unit 15, verification The processing unit 16 and the notification unit 17 are realized. Note that the program may be provided in a state of being recorded on a recording medium such as a CD-ROM, a DVD-ROM, or a memory card.

[Verification operation by verification device]
Next, the collation operation by the collation apparatus 1 according to the present embodiment will be described.
FIG. 2 is a diagram showing an overall processing flow of the collating operation by the collating device 1 according to the present embodiment.

  First, the image data acquisition unit 10 acquires image data 21 input from the outside (imaging device 2) (S1). For example, in the case of entrance / exit management, an image of a person who is going to enter a room is captured by a stereo camera or the like provided at the entrance, and the image data of the image is input to the image data acquisition unit 10. At this time, the imaging device 2 captures at least two images having a difference in photographing time. For example, the photographing timing may be shifted twice, and the image data obtained by each photographing may be input to the image data acquisition unit 10, or the imaging device 2 may be a stereo camera capable of photographing an interlaced image. In some cases, odd-numbered frame image data and even-numbered frame image data of an interlaced image obtained by one shooting may be input to the image data acquisition unit 10. The image data 21 acquired by the image data acquisition unit 10 is temporarily stored in the storage unit 15, for example.

  Next, the area detection unit 11 reads the image data 21 acquired in step S1 from, for example, the storage unit 15, and detects an area including a characteristic portion included in the image data 21 (S2). For example, the area detection unit 11 is specified based on two image data acquired by the image data acquisition unit 10 by a face area of a subject (person) included in each image data, two pupils, and a nose. Each characteristic area is detected, and area information of the detected area is generated.

Thereafter, the motion detection processing unit 12 performs a motion detection process based on the image data 21 acquired in step S1 and the area information of the area detected in step S2 (S3). Specific processing contents of the motion detection processing will be described later.
As a result of the motion detection process in step S3, when it is determined that the subject of the image data 21 acquired in step S1 has a motion, the notification unit 17 notifies the user of an error and also the imaging device 2. Is instructed to re-shoot (S4). Then, it returns to step S1 again.

  On the other hand, if it is determined as a result of the motion detection process in step S3 that the subject of the image data 21 acquired in step S1 has no motion, the collation data generation unit 13 is based on the image data 21 acquired in step S1. Then, three-dimensional measurement is performed (S5). For example, the collation data generation unit 13 performs three-dimensional measurement based on the two-dimensional image data and the three-dimensional image data included in any one of the two image data 21 acquired in step S1. Next, the collation data generation part 13 produces | generates collation data based on the measurement result of the three-dimensional measurement performed at step S5 (S6).

When the collation data is generated, the collation processing unit 16 performs a collation process for collating the collation data generated in step S6 with the registered image data 151 stored in the storage unit 15, and the collation data is any one. A determination result as to whether or not it matches the registered image data 151 is output (S7). For example, in entry / exit management, when it is determined that the collation data matches any of the registered image data 151, the collation apparatus 1 performs control to open the entrance / exit door. On the other hand, if it is determined that the collation data does not match any registered image data 151, the collation device 1 displays error information on the screen of the console, for example, without opening the entrance / exit door, By making a sound indicating an error from the speaker, the user is notified that the identity verification has failed.
Thus, a series of processes at the time of collation is completed.

[Motion detection processing]
Next, the motion detection process in step S3 will be described in detail.
FIG. 3 is a diagram illustrating a process flow of the motion detection process. Here, a case where image data of an interlaced image is input in step S1 will be described as an example.

  First, two pieces of image data to be subjected to motion detection processing are input to the motion detection processing unit 12 (S31). Specifically, two pieces of image data (image data of an odd frame and an even frame of an interlaced image) having a photographing time difference acquired in step S1 are input.

Next, the motion detection processing unit 12 extracts the image data of the area detected by the area detection unit 11 from the two input image data (S32).
FIG. 4 shows an example of extraction of feature area image data by the motion detection processing unit 12. In the figure, reference numeral 200 indicates the interlaced image acquired in step S1, reference numeral 201 indicates the face area in the interlaced image 200 detected by the area detection unit 11 in step S2, and reference numeral 202 indicates The characteristic area in the face area 201 of the interlaced image 200 detected by the area detection unit 11 in step S2 is shown. In the figure, the case where the area specified by the two pupils and nose of the person is the feature area 202 is shown as an example.
As shown in FIG. 4, the motion detection processing unit 12 is specified by area information (coordinate information) generated by the area detection unit 11 from, for example, odd frame image data and even frame image data of the interlaced image 200. The image data 203 and 204 of the area 202 to be extracted are extracted.

Next, the motion detection processing unit 12 calculates a correlation between the two image data extracted in step S32 (S33). Specifically, the motion detection processing unit 12 uses the phase only correlation method based on the two image data 203 and 204 extracted in step S32, and the movement amounts dx and dy and the correlation between the two image data 203 and 204. The intensity a is calculated. More specifically, the motion detection processing unit 12 first generates image data obtained by thinning out pixels from the two image data 203 and 204 extracted in step S32. For example, 64 × 64 pixel image data is generated by thinning out two pixels alternately in the horizontal direction (row direction) in the two image data 203 and 204. Next, movement amounts dx, dy and correlation strength a between two pieces of image data generated by thinning out pixels are calculated by the phase-only correlation method.
5A and 5B show an example of correlation characteristics between two pieces of image data by the phase only correlation method. FIG. 5A shows the correlation characteristics when there is a difference between two image data (the image is blurred), and FIG. 5B shows the correlation when there is no difference between the two image data (the image is not blurred). Show properties.

  When there is a difference between the two image data, as shown in FIG. 5A, the movement amounts dx and dy in the x direction and the y direction increase, and the correlation strength a (peak value peak) decreases. On the other hand, when there is no difference between the two image data, as shown in FIG. 5B, the movement amounts dx and dy in the x direction and the y direction are reduced, and the correlation strength a is increased.

Next, the motion detection processing unit 12 compares the movement amounts dx, dy and the correlation strength a calculated in step S34 with predetermined threshold values (S34). For example, the threshold value xth of the movement amount dx, the threshold value yth of the movement amount dy, and the threshold value ath of the correlation strength a are set in advance. The motion detection processing unit 12 compares the calculated movement amount dx with the threshold value xth, compares the calculated movement amount dy with the threshold value yth, and compares the calculated correlation strength a with the threshold value ath.
The motion detection processing unit 12 performs the case where the movement amount dx is smaller than the threshold value xth, the movement amount dx is smaller than the threshold value xth, and the correlation strength a is larger than the threshold value ath (dx <xth, dy <yth, a> ath). Then, it is determined that there is no movement of the subject included in the image data (S35). On the other hand, the motion detection processing unit 12 satisfies at least one of the following conditions: the movement amount dx is larger than the threshold value xth, the movement amount dx is larger than the threshold value xth, and the correlation strength a is smaller than the threshold value ath ( If the condition is other than dx <xth, dy <yth, and a> ath), it is determined that there is a movement of the subject included in the image data (S36). For example, if the threshold values are set as | xth | = 1, | yth | = 1, and ath = 0.4, in the case of FIG. 5A, the movement amount dx = −5, the movement amount dy = 4, Since the correlation strength a = 0.28, it is determined that there is a movement of the subject included in the image data. On the other hand, in the case of FIG. 5B, since the movement amount dx = 0, the movement amount dy = 0, and the correlation strength a = 0.89, it is determined that there is no movement of the subject included in the image data.
As described above, a series of the motion detection processing ends.

[Verification processing]
Next, the collation process in step S7 will be described in detail.
FIG. 6 is a diagram illustrating a processing flow of the collation processing. Here, a case where image data of an interlaced image is input in step S1 will be described as an example.

  First, the collation processing unit 16 initializes values of parameters used for the collation processing (S71). Examples of the parameters include a verified image number t indicating the number of registered images (data) that have been verified, and a registered image number N that defines the number of registered images (data) to be verified. Specifically, the collation processing unit 16 sets the number of registered image data 151 stored in the storage unit 50 as the registered image number N, and sets the collated image number t to “0 (zero)”.

  Next, the collation processing unit 16 determines whether or not the collated image number t is smaller than the registered image number N (S72). When the collated image number t is smaller than the registered image number N, the collation processing unit 16 reads the registered image data 151 to which the t-th serial number is assigned from the storage unit 15. For example, when the number of collated images t = 1, registered image data 151_1 to which the reference number “1” is assigned is read from the storage unit 15.

  The matching processing unit 16 compares the registered image data 151 read in step S73 with the matching data generated in step S6, and calculates a matching score (S74). Here, the collation score is a value indicating the degree of similarity between the collation data and the registered image data. For example, the matching processing unit 16 searches for a feature area of the matching data by the phase-only correlation method, compares this feature area with an area in the registered image data corresponding to the feature area, and both are similar. The degree is calculated as a matching score. In the present embodiment, it is assumed that the collation score is more similar as the numerical value is higher.

  When the collation score is calculated, the collation processing unit 16 increments the numerical value of the collated image number t (S75). Thereafter, the process returns to step S72 again, and the processes of steps S72 to S75 are repeatedly executed until the collated image number t becomes larger than the registered image number N. Thereby, the collation score for each registered image data registered in the storage unit 15 is calculated.

  If the collated image number t is larger than the registered image number N in step S72, the collation processing unit 16 executes determination processing based on the collation score for each registered image data calculated in step S74 (S76). . For example, the matching processing unit 16 selects the registered image data having the highest matching score from the registered image data for which the matching score has been calculated in step S74, and matches the matching data when the selected matching score exceeds the threshold value. It is determined that there is registered image data to be registered, and if the selected collation score does not exceed the threshold value, it is determined that there is no registered image data. Thus, a series of collation processes is completed.

[Image registration by collator]
Next, an image registration operation by the collation device according to the present embodiment will be described.
FIG. 7 is a diagram showing an overall processing flow of the image registration operation by the collation apparatus according to the present embodiment.
For example, in the case of entrance / exit management, first, when an administrator or the like operates a console installed in the vicinity of a stereo camera or the like provided at the entrance, the registration of an image used for personal verification is instructed. The apparatus 1 starts an image registration operation. The processing after the start of the image registration operation is the same as the above-described collation operation (FIG. 2) from step S1 to S6. In step S <b> 6, when collation data related to the image data acquired in step S <b> 1 is generated, the image registration unit 14 stores the generated collation data in the storage unit 15 as registered image data 151. At this time, the image registration unit 14 assigns a reference number different from the reference number assigned to the registered image data 151 already stored in the storage unit 15, and stores the registered image data in the storage unit 15 together with the reference number. To do. Thus, a series of operations for image registration is completed.

  As described above, according to the collation device according to the present embodiment, at the time of collation, the presence / absence of motion of the subject included in the image captured by the imaging device (presence / absence of image blur) is detected by the motion detection process. In addition, since it is possible to collate images when it is determined that there is no movement of the subject, it is possible to acquire only images without blurring, and to prevent collation processing based on blurred images. As a result, it becomes difficult for the verification of the identity verification or misidentification with others to occur, and the accuracy of verification can be improved. In addition, when it is determined that there is movement of the subject by the motion detection process, image verification is not performed, so it is possible to omit useless processing that has been performed in the case of conventional verification failure, and overall processing related to verification Time can be shortened.

  In addition, when an image is registered, the image is registered only when it is determined that there is no movement of the subject. Therefore, an image without blur can be registered as a registered image, which contributes to improvement of collation accuracy.

  Further, according to the collation device according to the present embodiment, since the presence / absence of the movement of the subject is determined based on the movement amounts dx, dy calculated by the phase-only correlation method and the correlation strength a in the motion detection process, high detection is performed. Accuracy can be expected. In addition, since the odd frame image data and the even frame image data of the interlaced image are used as the two image data used in the phase only correlation method, it is only necessary to perform shooting once by the imaging device in order to obtain the two image data. This saves you the trouble of shooting.

  Furthermore, according to the collation device according to the present embodiment, when it is determined that there is a movement of the subject by the motion detection process, an error is notified to the user and a re-shooting is instructed to the imaging device. It is possible to reduce the time until re-photographing for redoing, and further reduce the overall processing time required for collation.

  Although the invention made by the present inventors has been specifically described based on the embodiments, it is needless to say that the present invention is not limited thereto and can be variously modified without departing from the gist thereof. Yes.

  For example, in the above embodiment, a case where an area specified from two pupils and a nose is extracted as a feature area of a face image is exemplified. However, a feature part of a face (eye, nose, mouth, eyebrows, etc.) is extracted. Other areas may be used as long as they are included.

  In the above embodiment, the case where the data of the feature area 202 extracted from the image data 200 acquired by the image acquisition unit 10 is extracted to calculate the movement amount and the like is illustrated (see FIG. 4). The movement amount or the like may be calculated by extracting the data, or the movement amount or the like may be calculated from the entire image data 200 without extracting the feature area.

  In the above embodiment, the case where the collation device registers and collates the image data of the face image is exemplified. However, the same applies to the collation device that registers and collates image data such as a fingerprint instead of the face image. be able to.

  Also, in the above embodiment, the case where the presence or absence of motion of the acquired face image is detected and a technique for performing various processes based on the acquired face image when it is determined that there is no motion is exemplified in the face authentication device However, the present invention is not limited to this, and the above technique can be applied to, for example, a surveillance camera. For example, it is possible to detect the presence or absence of the movement of the subject from the image data captured by the monitoring camera, and to search only the face image without the movement of the subject.

  DESCRIPTION OF SYMBOLS 100 ... Authentication system, 1 ... Collation apparatus, 2 ... Imaging device, 10 ... Image data acquisition part, 11 ... Area detection part, 12 ... Motion detection process part, 13 ... Collation data generation part, 14 ... Image registration part, 15 ... Storage unit, 16 ... collation processing unit, 17 ... notification unit, 21, 21_1, 21_2 ... image data, 21, 21_1, 21_2 ... image data, 151_1 to 151_n ... registered image data, 200 ... image data (interlaced image), 201 ... Face area, 202 ... Feature area, 203 ... Image data of odd frames, 204 ... Image data of even frames.

Claims (11)

An image data acquisition unit for acquiring image data of an image captured by the imaging device;
Based on the image data acquired by the image data acquisition unit, a motion detection processing unit for detecting the presence or absence of movement of the subject included in the image data;
When the movement of the subject is not detected by the motion detection processing unit, verification data used for verification is generated based on the image data acquired by the image data acquisition unit, and the motion detection processing unit A collation data generation unit that does not generate the collation data when a motion is detected;
A storage unit for storing registered image data in which the subject to be verified is copied;
A collation apparatus comprising: a collation processing unit that collates the collation data generated by the collation data generation unit and the registered image data stored in the storage unit. The verification apparatus according to claim 1,
The collation apparatus further comprising an image registration unit that stores the collation data generated by the collation data generation unit in the storage unit as the registered image data. The verification apparatus according to claim 1,
The motion detection processing unit calculates a movement amount and correlation strength of the two image data based on two image data having a photographing time difference acquired by the image data acquisition unit by a phase only correlation method, and calculates the calculated A collation apparatus that detects presence or absence of movement of the subject based on a movement amount and the correlation strength. In the collation device according to claim 3,
An area detection unit for extracting an area including a characteristic part of the subject included in the image data acquired by the image data acquisition unit;
The motion detection processing unit extracts the data in the area detected by the area detection unit from the two image data, respectively, and based on the extracted data, the movement amount between the two image data and the A collation device that calculates correlation strength. In the collation device according to claim 3 or 4,
The two image data are odd frame data and even frame data of an interlaced image. In the collation device according to any one of claims 1 to 5,
The collating apparatus further comprising: a notifying unit that notifies an error and instructs the imaging device to perform re-shooting when the motion of the subject is detected by the motion detection processing unit. A first step of acquiring image data of an image photographed by the imaging device;
A second step of detecting presence or absence of movement of a subject included in the image data based on the image data acquired by the image data acquisition unit;
A third step of generating collation data used for collation based on the image data acquired by the image data acquisition unit when the movement of the subject is not detected in the second step;
And a fourth step of collating the collation data generated in the third step with registered image data in which the subject to be collated stored in the storage unit is copied. The collation method according to claim 7,
The second step includes
A fifth step of calculating a movement amount and correlation strength of the two image data based on two image data having a photographing time difference obtained in the first step by a phase only correlation method, and the calculated movement amount and the And a sixth step of detecting the presence / absence of the movement of the subject based on the correlation strength. The verification method according to claim 8,
A seventh step of detecting an area including a characteristic portion of the subject included in the image data acquired by the first step;
The fifth step is a step of extracting the area data detected in the seventh step from the two image data, and calculating the movement amount and the correlation strength based on the extracted data. A matching method characterized by In the collation method according to claim 8 or 9,
The two image data are odd frame data and even frame data of an interlaced image. In the collation method as described in any one of Claims 7 thru | or 10,
A verification method further comprising an eighth step of notifying an error and instructing the imaging apparatus to perform re-imaging when a movement of the subject is detected in the second step.